Wireless Site Surveys: Predictive, Passive and Active Surveys Compared

A wireless site survey determines where access points should go and confirms that the result works. Skipping it is why so many wireless networks show full signal bars and perform badly — coverage is easy to achieve and is not the same thing as a network that works, which is the single most useful idea in this topic.

The three survey types

Predictive surveys are done in software from floor plans. You model the building, assign attenuation values to walls and materials, and the tool proposes access point placement and predicts coverage. It is fast, requires no site visit, and it is only as accurate as the model — a concrete wall entered as drywall produces a confident and wrong answer. Use it for initial design, then verify on site.

Passive surveys walk the site with a device listening to what is already broadcasting, recording signal strength, noise, and channel usage at each point without associating to anything. This is how you measure existing coverage, find dead spots, discover neighbouring networks competing for spectrum, and identify rogue access points. It measures what the air looks like.

Active surveys associate to the network and generate real traffic, measuring throughput, latency, packet loss, and roaming behaviour. This is what tells you whether the network actually works, because a location can show strong signal and still deliver poor performance from interference, contention, or a misconfigured access point.

The standard sequence is predictive to design, passive to verify coverage after installation, active to validate performance, then a repeat survey after any significant change to the building or the client population.

Reading the measurements

Signal strength is reported as RSSI in negative dBm, where closer to zero is stronger. Roughly, above −67 dBm supports voice and video, around −70 is adequate for general data, and below −80 is unreliable. Design to the requirement rather than to the best number achievable.

Signal-to-noise ratio is often more telling than raw signal. A strong signal in a noisy environment performs worse than a moderate signal in a quiet one, and 25 dB of SNR is a common target for reliable throughput.

Channel utilization reveals contention. Wireless is a shared medium, so a channel already busy with neighbouring networks delivers poor throughput regardless of how strong your signal is. Co-channel interference — your own access points sharing a channel and taking turns — is a self-inflicted version of the same problem.

Understanding which identifiers you are looking at helps here: the network name, the per-radio identifier, and the extended service set are distinct things, as covered under SSID, BSSID and ESSID, and a survey tool reports per-BSSID.

Coverage versus capacity

A coverage-driven design asks how few access points can reach every corner. A capacity-driven design asks how many clients will be in each area and how much each needs.

Modern deployments are almost always capacity-driven. A lecture hall or open-plan floor may need several access points within sight of each other, each at reduced transmit power so their cells stay small and clients distribute across them. Turning power up to cover more area with fewer units is the classic mistake: it enlarges cells, increases co-channel interference, and encourages distant clients to cling to a far access point at low data rates, slowing the cell for everyone.

Channel planning follows. The 2.4 GHz band has only three non-overlapping channels and is best treated as legacy support; 5 GHz has many more; 6 GHz more still where clients support it. Reuse channels with enough physical separation that cells on the same channel do not overlap.

Practical considerations

Survey with the devices that will actually be used, because a laptop with a good antenna hears access points a handheld scanner or an IoT sensor cannot. Design for the weakest client, not the best one.

Survey under realistic conditions — an empty warehouse and a full one behave differently, and people, stock, and vehicles all absorb radio energy. Account for what is in walls and ceilings: metal, concrete, water pipes, elevator shafts, and reflective glass.

Finally, check the wired side. Access points are powered over the cable, so verify the switch's power budget and standard support before installing, as covered under Power over Ethernet — a perfect wireless design fails if the closet cannot power it.

Exam angles

Know predictive (software model), passive (listen without associating), and active (associate and test) surveys and when each is used. Understand that a heat map shows signal strength and that strong signal does not guarantee performance. Recognize channel utilization and signal-to-noise ratio as the metrics that explain poor throughput despite good coverage, and co-channel interference and excessive transmit power as common design faults.